Stefan Petri's Publications

Some LaTeX Macros to begin with

%
% THE \simulan LOGO IS DEFINED HERE.
%
\def\simulan{{\rm s\kern-.06em\raise-.5ex\hbox{i}\kern-.1em\raise-.1ex
    \hbox{m}\raise-.3ex\hbox{u}\kern-.10emL\kern-.1667em\lower-.6ex
    \hbox{a}\kern-.10emn}}

%% the \PBeam Logo is defined here
\def\PBeam{{\sc\kern.15emP\kern-.9em\raise.125ex\hbox{$\leftarrow$}\sc\kern-.25emB\sc\kern-.1eme\kern-.1ema\kern-.1emm}}

Diploma Thesis

[Petri91] Stefan Petri:
\simulan - Ein Werkzeug zur Simulation lokaler Netze.
Diplomarbeit, TU Braunschweig, 1991 (in german).
Abstract: With decreasing hardware costs and increasing computing power the interconnection of Personal Computers, Workstations and special purpose hardware becomes more and more important. \simulan{} is a tool for efficient and flexible modelling and simulation of such networks. Models for new protocols and applications can be programmed with little effort. The concepts of \simuLan{} are shown, the usage of the program is explained and the possibilities for customizing are discussed.
You might also want look at the companion thesis about modelling with simuLan:
[Schmida91] Ralf Schmidt-Dannert:
\simulan: Modellierung und Simulation lokaler Netzwerke.
Diplomarbeit, TU Braunschweig, 1991 (in german).

Dissertation

[Petri97Book] Stefan Petri:
Lastausgleich und Fehlertoleranz in Workstation-Clustern.
PhD Thesis, Institut für Betriebssysteme und Rechnerverbund, TU Braunschweig, Mai 1997 (in german).
Shaker Verlag, Aachen, ISBN 3-8265-2471-3.
Abstract: Networked workstations are commonly used in companies and research institutes to provide users with interactive computing facilities. It was discovered early that the processing capacities of such installations typically show a low utilization. These workstation clusters are an increasingly attractive platform for resource intensive distributed and parallel applications because of their good price-performance ratio.
However, they exhibit problems with unbalanced load, conflicts with interactive users and increased failure probability. Process migration and checkpoint/restart are adequate means to solve these problems. Several of such mechanisms are described in the literature, but they all have certain weaknesses and impose restrictions on the applications they can handle.
In this thesis, a new concept for an application transparent migration and checkpointing mechanism is developed, wich overcomes some substantial of those restrictions. It supports migration and fault transparency for parallel and distributed applications, i.e. groups of communicating processes, on clusters of workstations. Neither the system kernel nor the application programs need to be modified, and applications are not required to be written for a specific runtime environment. However, for better performance the applications can be linked with a modified system library. From this concept, the architecture of the example implementation $\PBeam$ is derived, and first measured results and experiences with the implementation are reported.
A BiBTeX-File, from my Dissertation.

Two Journal Articles

[PL95] Stefan Petri and Horst Langendörfer:
Load Balancing and Fault Tolerance in Workstation Clusters - Migrating Groups of Communicating Processes.
Operating Systems Review, Vol. 29, No. 4, October 1995, pp. 25-36.
Annote:The first publication about the general concept (virtual name space) and very first experiences with the single process checkpointer. Leaves many questions about checkpointing groups of processes open, which are solved in subsequent work.
Abstract: In the past, several process migration facilities for distributed systems have been developed. Due to the complex nature of the subject, all those facilities have limitations that make them usable for only limited classes of applications and environments. We discuss some of the usual limitations and possible solutions. Specifically, we focus on migration of groups of collaborating processes between Unix systems without kernel modifications, and from this we derive the design for a migration system. First experiences with our implementation show that we reach performance figures for the migration that are close to those of real distributed operating systems.
[Petri98] Stefan Petri:
Lastausgleichswerkzeuge für Workstation-Cluster.
Informatik/Informatique -- Zeitschrift der schweizerischen Informatikerorganisationen / Revue des organisations suisses d'informatique, Themenheft ``Workstations'', No. 3, June 1998, pp. 23-26.
Abstract: Programmierer und Benutzer rechenintensiver verteilter Anwendungen wollen oder gar können Lastausgleich und Fehlertoleranz nicht in ihren Programmen behandeln. Da die verfügbaren Betriebssysteme solche Funktionalitäten nur rudimentär bieten, werden in diesem Artikel zwei anwendungstransparente Migrationssysteme, $\PBeam$ und CoCheck, vorgestellt und verglichen. Abschließend wird ein Ausblick auf neue Funktionalitäten gegeben, die durch Integration mit interaktiven Werkzeugen erreicht werden können.

Conference Papers

[PSL96] Stefan Petri and Bettina Schnor and Horst Langendörfer:
$\PBeam$ -- Fehlertoleranz für verteilte Anwendungen mittels Migration und Checkpointing.
In Clemens H. Cap (Ed.) Workstations und ihre Anwendungen - Proceedings der Fachtagung SIWORK'96, May 14./15. 1996, vdf Hochschulverlag, Zürich, pages 91-102 (in german).
Abstract: Process migration and checkpointing offer the possibility to achieve load balancing and fault tolerance. We discuss some of the problems and possible solutions for supporting these for distributed applications in workstation clusters. We describe the design for our $\PBeam$ system and its facility for creating global consistent checkpoints of groups of communication process on networked Unix systems, and present first experiences with our implementation.
[PSLS96] Stefan Petri and Bettina Schnor and Horst Langendörfer and Jens Steinborn:
Consistent Global Checkpoints for Distributed Applications on Clusters of Workstations.
In Hermann G. Matthies und Josef Schüle (Eds.), Paralleles und Verteiltes Rechnen - Beiträge zum 4. Workshop über Wissenschaftliches Rechnen, pages 77-86, Braunschweig, 9.-11. October 1996. Shaker Verlag.
Annote: Continuation of the SIWORK paper, updated with newer results.
Abstract: Process migration and checkpointing offer the possibility to achieve load balancing and fault tolerance. We give an overview of the design of our $\PBeam$ system and focus on its facility for creating global consistent checkpoints of groups of communicating processes on networked Unix systems. We present first experiences with our implementation.
[PL96] Stefan Petri and Horst Langendörfer:
CURE -- Computer Assisted Rehabilitation.
In J. Klaus and E. Auff and W. Kremser and W.L. Zagler (Eds.): Interdisciplinary Aspects on Computers Helping People with Special Needs: 5th International Conference / ICCHP'96, pages 591-598, July, 17.-19. 1996, Linz. Oldenbourg Verlag.
Abstract: Specific training of the disordered functions can yield good success in the treatment of patients with injured central nervous system. To assist this staff-intensive treatment, computers, at which patients work on therapeutic exercises, have been in use for several years. After a brief summary of the experiences with an existing implementation NIXAS-CURE, we focus on the design and implementation of the new CURE (Computerunterstützte Rehabilitation = computer assisted rehabilitation) version 2 system that is developed by Siemens-Nixdorf and TU Braunschweig. This version exploits modern PCs' facilities to create user interfaces that are suited for handicapped persons. Furthermore, it gives the therapist improved support for planning and evaluating the therapy course.
[PSBHTL97] Stefan Petri and Bettina Schnor and Matthias Becker and Bernd Hinrichs and Titus Tscharntke and Horst Langendörfer:
Evaluation of Multicast Methods to Maintain a Global Name Space for Transparent Process Migration in Workstation Clusters.
In Martina Zitterbart (Ed.), Kommunikation in Verteilten Systemen, GI/ITG Fachtagung KIVS'97, pages 224-234, Braunschweig, 18.-21. February 1997. Springer Verlag.
Abstract: In this paper we examine the performance and fault tolerance of different multicast based methods for maintaining the consistence of distributed data, depending on the network topology and packet loss probability. We present a system model and analytical results. This work is motivated by experiences with our process migration system $\PBeam$ . There we use a global virtual name space to achieve location transparency for process migration and checkpointing~/ rollback for distributed applications on clusters of Unix workstations. First measurements have shown that the maintenance of this global name space is critical for the performance of the entire system.
[PBL98] Stefan Petri and Matthias Bolz and Horst Langendörfer:
Migration and Rollback Transparency for arbitrary Distributed Applications in Workstation Clusters.
In Jose Rolim (Ed.), Parallel and Distributed Processing - 10 IPPS/SDPD'98 Workshops held in conjunction with the 12th International Parallel Processing Symposium and 9th Symposium on Parallel and Distributed Processing, Orlando, Florida, USA; Proceeedings, 30. March 1998, Orlando, Florida. Lecture Notes in Computer Science, Volume 1388, pp. 159-170, Berlin, March 1998. Springer.
ISI:000077612600023
2nd Workshop on Runtime Support for Parallel Programming (RTSPP); extended version published as technical report TU Braunschweig, Informatik-Bericht 98-02.
Abstract: Programmers and users of compute intensive scientific applications often do not want to (or even cannot) code load balancing and fault tolerance into their programs.
The $\PBeam$ system uses a global virtual name space to provide migration and rollback transparency in user space for distributed groups of processes on workstations. Applications always use the same virtual names for the operating system objects, independent of their current real location. The system calls are interposed and their parameters translated between the name spaces. Unlike other migration mechanisms, $\PBeam$ does not require the applications to be written for a specific programming model or communication library.
The first approach to execute applications in the virtual name space was to link the programs with a modified system library. Now, in this paper we describe design and implementation of a separate system call interposition process that accesses the application via the debugging interface. The main advantage of this approach is that it can handle even unmodified (e.g. commercially bought) application programs. We compare measured performance figures with previous similar approaches and the modified system library.
[PLGHOB99] Stefan Petri and Gunther Lustig and Claus Grewe and Rainer Hagenau and Wolfgang Obelöer and Marcel Boosten:
Performance Comparison of different High-Speed Networks with a Uniform Efficient Programming Interface.
In: Wolfgang Karl and Geir Horn (Eds), Scalable Coherent Interface -- Proceedings of SCI Europe '99, September 2nd - 3rd, 1999, pages 83 - 90, held as a Conference Stream of Euro-PAR'99, Toulouse, France, 31 Aug. - 3 Sept. 1999.
Annote: Continuation of the CC'99 contribution.
Abstract: The communication network and its interface are an important factor for the performance of modern cluster computing systems. During the last years, several new approaches were developed which yield higher bandwidth and lower latency than traditional alternatives. However, applications often can exploit the performance of the underlying network only by using specially tuned programming interfaces. This dependency makes porting the applications difficult if not impossible. Therefore, with HPCC (High Performance Cluster Communication) we have designed a programming interface which enables applications to use different high speed networks in a uniform and efficient manner. Currently it is implemented for four high speed networks: HIC, SCI, Gigabit Ethernet, and Myrinet. This paper gives a short overview of the HPCC programming interface, and presents comparative figures for the performance of basic operations on four different high speed network technologies.
[Petri99] Stefan Petri:
A Common Framework for Transparent Checkpointing, Replication and Migration in Clusters.
In: W. Erhard, K.-E. Großpietsch, W. Koch, E. Maehle and E. Zehendner, (Eds.), ARCS'99: Architektur von Rechensystemen, Vorträge der Workshops im Rahmen der 15. GI/ITG-Fachtagung vom 4. bis 7. Oktober 1999 in Jena pp. 79-88, Jena, October 1999.
Workshop Verläßlichkeit und Fehlertoleranz.
Abstract: This paper presents a new approach to unify fault tolerance and load balancing in a common framework, aimed at providing those services (even) to legacy applications, on clusters of off-the-shelf PCs or workstations. Such clusters with high speed networks are attractive due to their price-performance ratio and rising power. However, multi-user operation and high node failure probability necessitate resource management and fault tolerance measures for long running applications, and for applications that require high availability. Now, programmers and users of applications are interested in their algorithms, solutions or services, and expect resource management and fault tolerance as services from the run time environment. Unfortunately, such clusters provide these services only minimally, if at all. A suitable approach to solve this discrepancy is application transparent checkpointing, migration, rollback and replication. It should be implemented as middleware between operating system and application to facilitate implementation and installation. We describe our work to extend our checkpointing and migration system $\PBeam$ to also provide replication for distributed applications. The approach to achieve these services transparently for the application is the construction of a system-wide virtual name space, which abstracts the running application processes from the actual location and time of their execution. No changes to the system kernel or the application code are required. Checkpointing already is the common base for rollback and for migration, and now we extend the migration to cloning to also achieve replication. With application transparent checkpoint / rollback, $\PBeam$ can only detect and treat crash faults. With replication it now is also able to detect other faults by comparing the output of the replica, even though it cannot know the meaning of the output contents. The abstraction through the virtual name space is an important prequisite to enable error detection through output comparison on cluster systems. It ensures that no environment specific information can yield differences in the replica outputs during fault-free operation.

Co-Author of some more Conference Papers

[SPL94] Jürgen Schöwälder and Stefan Petri and Horst Langendörfer:
Frei erhältliche Software für UNIX Administratoren.
In Offene Systeme -- Unix in Deutschland, GUUG Jahrestagung, Wiesbaden, September 1994, pages 322-329 (in german).
Abstract: The Internet provides a lot of free software for UNIX system administration. This survey of some of the most important programs should help new system administrators to find software that solves their problems. We further would like to motivate an open discussion about experiences with free software for system administration tasks.
[SLP94] Bettina Schnor and Horst Langendörfer and Stefan Petri:
Einsatz neuronaler Netze zur Lastbalancierung in Workstationclustern.
In: Horst Langendörfer (Ed.), Praxisorientierte Parallelverarbeitung - Beiträge zum 3. Workshop über Wissenschaftliches Rechnen, Hanser, München, October 1994, pages 154-165 (in german).
Abstract: ...
[SPL96] Bettina Schnor and Stefan Petri and Horst Langendörfer:
Load Management for Load Balancing on Heterogeneous Platforms: A Comparison of Traditional and Neural Network Based Approaches.
In: Luc Bouge, Pierre Fraigniaud, Anne Mignotte, Yves Robert (Eds.) Euro-Par'96 -- Parallel Processing Proceedings of the Second International Euro-Par Conference, Volume II, Springer Lecture Notes in Computer Science 1124, pages 615-620. ENS Lyon, France, August 27-29, 1996.
Abstract: In this paper we compare simple load metrics with neural networks which have been trained to predict the expected delay of an application from the sampled load informations. The results show that the proposed load metric performs well in heterogeneous environments. Further, neural networks can improve the performance of load balancing facilities.
[SPOL96] Bettina Schnor and Stefan Petri and Reinhard Oleyniczak and Horst Langendörfer:
Scheduling of Parallel Applications on Heterogeneous Workstation Clusters.
In: Koukou Yetongnon and Salim Hariri (Ed.), Proceedings of PDCS'96 -- ISCA 9th International Conference on Parallel and Distributed Computing Systems, pages 330-337, Dijon, France, September 1996.
Abstract: Scheduling and load balancing for parallel applications can be done at application level, or at system level. Application level scheduling is less transparent, because it has to be coded into each program, and can lead to contradicting decisions, when the different applications do not know about each other. Therefore, in this paper we focus on system level load balancing. We discuss the special properties of heterogeneous workstation clusters for scheduling parallel applications. A Shortest-Expected-Delay mapping is presented that assigns the processes of a parallel application to ``virtually homogeneous'' machines, based upon the current load situation. We present simulation results that show, how and when process migration is beneficial in heterogeneous systems.
[SPL96b] Bettina Schnor and Stefan Petri and Horst Langendörfer:
A Detailed Comparison of Traditional and Neural Network Based Approaches for Load Metrics on Heterogeneous Platforms.
In: Hermann G. Matthies und Josef Schüle (Eds.), Paralleles und Verteiltes Rechnen -- Beiträge zum 4. Workshop über Wissenschaftliches Rechnen, pages 93-106, Braunschweig, 9.-11. October 1996. Shaker Verlag.
Annote: Extended, continued and updated version of the EuroPar'96 Paper.
Abstract: Adaptive and dynamic load balancing are challenging problems in the field of parallel computing. Therefore, a load management facility is desirable which gives up to date load informations which are also comparable for heterogeneous machines. In this paper we discuss several approaches for load metrics in heterogeneous systems. We compare simple load metrics with neural networks which have been trained to predict the expected delay of an application from the sampled load informations. The results show that the proposed load metric performs well in heterogeneous environments. Further, neural networks can improve the performance of load balancing facilities.
[SPB98] Bettina Schnor and Stefan Petri and Matthias Becker:
Scalability of Multicast Based Synchronization Methods.
In Proceedings of the 24th EUROMICRO Conference, August 25-27, 1998, Västerås, Sweden, pages 969-975. IEEE Computer Society, August 1998.
Annote: Continuation of the work presented in the KIVS'97 paper.
Abstract: Consistent maintenance of distributed data is important in application areas like groupware and for runtime support for parallel computing. We examine the performance of different multicast based methods for maintaining the consistency of distributed data depending on the network topology and concurrency.
Some amount of data is kept distributed or replicated on some or all nodes of a distributed system. At every moment, each instance that accesses this data must see the same information. Updates must be delivered ordered, reliably, and efficiently.
Our prototype software implements ordered, reliable multicasts on top of the unreliable IP broad- or multicast with three different methods (Master-Slave, Token Exchange on Demand, Totem Single Ring). This paper shows measurement results for the efficiency and scalability of the three methods in different topologies.
The measurements confirm earlier analytical results. Totem behaves well in large networks with many concurrent senders. The overhead of Token on Demand and of the Master-Slave algorithm is almost the same. Also we could not find an indication for the often-read opinion that the Master-Slave approach scales worse because of the central bottleneck.
[GLOP99] Claus Grewe and Gunther Lustig and Wolfgang Obelöer and Stefan Petri:
Leistungsvergleich verschiedener Netzwerke mit einer einheitlichen und effizienten Benutzerschnittstelle.
In: Wolfgang Rehm and Theo Ungerer (Eds.), Cluster Computing -- Tagungsband zum 2. Workshop, Karlsruhe, 25./26. March 1999, pp. 125-134. Chemnitzer Informatik-Berichte CSR-99-02, March 1999 (in german).
Abstract: Ein leistungsbestimmender Faktor moderner Clustersysteme ist das verwendete Kommunikationsnetzwerk. In den letzten Jahren wurden hier eine Reihe neuerer Ansätze entwickelt, die höhere Bandbreiten und niedrigere Latenzzeiten im Vergleich zu klassischen Alternativen aufweisen. Allerdings können Applikationen die höhere Leistungsfähigkeit des zugrundeliegenden Netzwerkes häufig nur durch die Benutzung speziell abgestimmter Schnittstellen erzielen. Dieses mindert aber i.allg. die Portierbarkeit der Anwendungen. Aus diesem Grund wurde eine Benutzerschnittstelle konzipiert, die es aktuell ermöglicht, vier verschiedene Hochgeschwindigkeitsnetzwerke (HIC, SCI, Gigabit Ethernet, Myrinet) einheitlich und effizient zu nutzen. In diesem Beitrag wird ein kurzer Überblick über diese HPCC-Benutzerschnittstelle gegeben und Meßwerte für die vier Hochgeschwindigkeitsnetze vorgestellt.
[SP99] Bettina Schnor and Stefan Petri:
Using On-line Job Monitoring for Dynamic Load Balancing of Parallel Applications.
In: Hamid R. Arabnia et al. (Eds.), Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications, PDPTA'99, pages 2561-2567, Las Vegas, NV, USA, June 28 - July 1, 1999.
Abstract: While the significance of the communication costs for the performance of parallel applications is well known, these costs are hardly considered in the known resource management systems for workstation clusters. In this paper we present the Cluster Search Algorithm (CSA) for dynamic remapping of parallel applications in clusters. The novelty of our approach is that the communication behavior of the application does not have to be known a priori. The decisions are made dynamically at run time according to the actually observed communication behavior of the application and available resources. We present first simulation results and an overview over an actual implementation.
[PPG00] Holger Pals, Stefan Petri und Claus Grewe:
FANTOMAS: Fault Tolerance for Mobile Agents in Clusters.
In: J. Rolim et al., (Eds.), Parallel and Distributed Processing -- Proceedings of 15 IPDPS 2000 Workshops, Cancun, Mexico. Lecture Notes in Computer Science, Vol. 1800, pages 1236-1247, Springer, May 2000.
ISI:000171729800155
Annual IEEE Workshop on Fault-Tolerant Distributed Systems (FTPDS2000).
Abstract:
[SSP03] Lars Schneidenbach, Bettina Schnor and Stefan Petri:
Architecture and Implementation of the Socket Interface on Top of GAMMA.
In: 28th Annual IEEE Conference on Local Computer Networks (LCN 2003), Workshop on High Speed Local Networks (HSLN 2003), October 21, 2003, Bonn/Königswinter., pages 528-536. IEEE Computer Society.
Abstract:
[NHPVWL09] Thomas Nocke, Ursula Heyder, Stefan Petri, Katrin Vohland, Markus Wrobel and Wolfgang Lucht:
Visualization of Biosphere Changes in the Context of Climate Change.
In: Volker Wohlgemuth (Ed.), Information Technology and Climate Change -- 2nd International Conference IT for empowerment, September 26/27th 2008, Berlin., pages 29-36. trafo Wissenschaftsverlag.
Abstract:
[RPF09] Claus Rachimow, Stefan Petri and Michael Flechsig:
Manageing Metadata in a Collaborative Earth System Research Environment.
In: Volker Wohlgemuth (Ed.), Information Technology and Climate Change -- 2nd International Conference IT for empowerment, September 26/27th 2008, Berlin., pages 47-54. trafo Wissenschaftsverlag.
Abstract:

Co-Editor of one Workshop Proceedings

[ALV98] Arndt Bode and Andreas Ganz and Claudia Gold and Stefan Petri and Niels Reimer and Björn Schiemann and Thomas Schnekenburger (Editors):
Anwendungsbezogene Lastverteilung - ALV'98. SFB-Bericht Nr. 342/01/98 A, TU München, Institut für Informatik, Bericht Nr. TUM-I9806, February 1998.

Technical Reports

[PBL98b] Stefan Petri and Matthias Bolz and Horst Langendörfer:
Transparent Migration and Rollback for Unmodified Applications in Workstation Clusters.
TU Braunschweig, Informatik-Bericht Nr. 98-02, April 1998.
Annote: Extended version of the RTSPP/IPPS'98 paper, with more technical details and performance figures.
Abstract: See the 2nd RTSPP/IPPS98 paper above.
[GOPB98] Claus Grewe and Wolfgang Obelöer and Stefan Petri and Marcel Boosten:
Network Interface Software Development for Medical Applications.
ESPRIT Project 20693 ``ARCHES'' Deliverable 3.4.1, Institute for Computer Engineering of the MU Lübeck, December 1998.
Annote:
Abstract:
[DLOP99] Andreas C. Döring, Gunther Lustig, Wolfgang Obelöer and Stefan Petri:
Implementation of Virtual Channels for HIC.
Informatik-Bericht A-99-07, Institute for Computer Engineering of the MU Lübeck, April 1999.
Presented as contribution to the IEEE 1355 Interconnect and Routing Workshop, 7th International Conference on High Performance Computing and Networking Europe -- HPCN'99, Amsterdam.
Annote:
Abstract:
[GHP99] Claus Grewe and Rainer Hagenau and Stefan Petri:
Medical Application Adaption and Performance Evaluation.
ESPRIT Project 20693 ``ARCHES'' Deliverable 3.4.3, Institute for Computer Engineering of the MU Lübeck, September 1999.
Annote:
Abstract:
[PG99] Stefan Petri and Claus Grewe:
A Fault-tolerant Approach for Mobile Agents
In: Jan Hlavi{\v{c}}ka and Erik Maehle and Andr{\'a}s Pataricza (Eds), Dependable Computing~-- EDCC-3, Third European Dependable Computing Conference, Prague, Czech Republic, September 15-17, 1999, Fast Abstracts.
Annote:
Abstract:
[HP99] Rainer Hagenau and Stefan Petri:
ARCHES II -- Application, Refinement and Consolidation of HIC Exploiting Standards.
ESPRIT Project 20693 Final Report, IAB-Nr. 55, Institute for Computer Engineering of the MU Lübeck, December 1999.
Annote:
Abstract:
[Petri02] Stefan Petri:
Clusterbasiertes Supercomputing für interdisziplinäre Anwender
White Paper, PanMedium, Potsdam, August 2002.
Annote:
Abstract:
[Petri07] Stefan Petri:
How to set up a C3-Grid Data Provider site?
C3-Grid Deliverable C3Grid-AP4/7-Tut2-001-Draft, Potsdam, February 2007.
[RP07] Claus Rachimow and Stefan Petri:
Web-Based GUI for Editing C3-Grid Metadata
C3-Grid Deliverable C3Grid-Metadataeditor-002-Draft, Potsdam, February 2007.

Seminarbände

I was involved in supervising several students' seminars, where we have edited collections of the texts and talks. All these are in german.

[PuV94] Stefan Petri, Bettina Schnor and Horst Langendörfer, Herausgeber.
Seminarband Paralleles und Verteiltes Rechnen, Institut für Betriebssysteme und Rechnerverbund, TU Braunschweig, Sommersemester 1994.
[PuV95] Stefan Petri and Horst Langendörfer, Herausgeber.
Seminarband Paralleles und Verteiltes Rechnen - Fehlersuche und Fehlertoleranz in verteilten Systemen, Institut für Betriebssysteme und Rechnerverbund, TU Braunschweig, Sommersemester 1995.
[PuV96] Stefan Petri and Horst Langendörfer, Herausgeber.
Seminarband Paralleles und Verteiltes Rechnen - Lastausgleich in Verteilten Systemen, Institut für Betriebssysteme und Rechnerverbund, TU Braunschweig, Sommersemester 1996.

Stefan Petri's Homepage